Any experts on Lithium Thionyl Chloride batteries?

[NashConv]

Hi everyone,

I’m designing an ultra-low power sensor node and I need a sanity check on my battery life estimation assumptions.

I am unable to do long-term empirical profiling due to schedule constraints, so I am trying to bridge the gap between the datasheet's "continuous load" curves and my actual "pulse" application.

The System:

Battery: Li/SOCl₂ Bobbin Cell (1500mAh nominal).

Temp Range: -20°C to +85°C.

Load Profile:

Sleep: ~56µA base.

Heartbeat: 10mA peak for 4ms (Every 1 second) - intended to keep battery depassivated.

TX: 40mA peak for 200ms (Every 20 minutes).

The Questions:

1. Pulse vs. Continuous Derating at -20°C Datasheets for bobbin cells typically show a massive capacity drop at -20°C (down to ~50-60% of nominal) when under continuous loads like 10-40mA. Since my 40mA load is only 200ms followed by a 20-minute rest (allowing ion diffusion/recovery), am I correct in assuming the effective capacity loss is much lower?

Assumption: I am estimating ~80-85% accessible capacity at -20°C for this pulse profile, rather than the datasheet's 60%. Is this realistic, or am I being too optimistic about the recovery effect?

2. Cold Capacity Reversibility If the device sits at -20°C for 24 hours and then warms up to +5°C, does the "lost" capacity return?

Assumption: The capacity loss at cold temps is purely due to viscosity/impedance (the "pipe" shrinking) and not permanent chemical loss (the "tank" shrinking). Therefore, if the battery warms up, that capacity becomes accessible again.

3. Passivation & Self-Discharge at +85°C We have a phase where the device sits at +85°C.

Passivation: Is my 10mA / 4ms pulse @ 1Hz sufficient to prevent voltage delay at these temps?

Leakage: I know self-discharge follows the Arrhenius rule. At +85°C, is it reasonable to model this as ~10-15% annualized loss?

Any insights from those who have deployed bobbin cells in similar environments would be appreciated.

Thanks!

[littlepea]

Sleep: ~56µA base.

Heartbeat: 10mA peak for 4ms (Every 1 second) - intended to keep battery depassivated.

TX: 40mA peak for 200ms (Every 20 minutes).


can you optimize above? Lora or lorawan?

[Gribo]

You don't need such a high de-passivation duty cycle. Once the cell is depassivated, 56uA should keep it so under normal temperature conditions.
Make sure your MCU can reliably wake up even when the cell has high ESR due to passivation - this way, you can selectively apply loads and get it out of passivation.